基于背景抑制的星空图像目标运动轨迹提取

研究了空间观测CCD图像弱小目标检测与运动轨迹提取技术，介绍了一种基于背景抑制的多目标检测与运动轨迹提取方法。首先，对图像进行阈值处理，滤除图像的背景灰度和噪声；然后生成基于序列图像叠加与形态学变换相结合的背景掩模图像，对序列图像进行屏蔽处理，得到运动目标图像；对目标图像进行交叉投影算法确定目标区域，并进一步精确确定目标质心，进而提取目标的运动轨迹。实验证明所提出的算法不仅能够有效检测星空序列图像中的多个弱信号小目标，而且具有较强的实时性。

Target trace acquisition method of star images based on background elimination

An investigation about an algorithm for space target trace acquisition is carriedout. An algorithm for multi-target trace acquisition method based on background elimination of CCD star images is put forward. First is the pretreatment, every pixel whose gray value is below the threshold will be set as a black pixel, other pixels remain invariable. Then, a mask is created based on addition of the serial images dealt by threshold previously and morphologic dilation operation. The mask can shield fixed stars and preserve the moving targets. In the following, a cross-projection method is used in target region determination, and the precise position of the targets can be calculated using centroid method from the target region, so the targets trace can be acquired. Experimental results show that the algorithm can detect faint targets effectively in serial star images with virtues of high real-time.